[prev in list] [next in list] [prev in thread] [next in thread] 

List:       pykde
Subject:    Re: [PyQt] Smartpointer to Python
From:       "Arve Knudsen" <arve.knudsen () gmail ! com>
Date:       2008-05-27 11:21:10
Message-ID: a0d6258d0805270421q63e81b42j50adf7e5a260c2d0 () mail ! gmail ! com
[Download RAW message or body]

On Mon, May 26, 2008 at 11:32 PM, Phil Thompson
<phil@riverbankcomputing.com> wrote:
> On Monday 26 May 2008 9:34:00 pm Arve Knudsen wrote:
> > On 5/26/08, Phil Thompson <phil@riverbankcomputing.com> wrote:
> > > On Monday 26 May 2008 2:10:53 pm Arve Knudsen wrote:
> > > > On Mon, May 26, 2008 at 2:37 PM, Phil Thompson
> > > > 
> > > > <phil@riverbankcomputing.com> wrote:
> > > > > On Monday 26 May 2008 1:10:11 pm Arve Knudsen wrote:
> > > > > > Hi
> > > > > > 
> > > > > > Reading the SIP documentation I can't find any info on how C++
> > > > > > objects are deleted (returned to the heap) from Python, only that
> > > > > > their destructor is called.
> > > > > 
> > > > > ...when the Python object is garbage collected and if the C++ object
> > > > > is owned by Python.
> > > > 
> > > > Yes, that much is clear.
> > > > 
> > > > > > Does this mean that when returning new'ed
> > > > > > pointers  to Python they should be managed by smartpointers? If so,
> > > > > > is there any mechanism in SIP for treating smartpointers as objects
> > > > > > of their managed type from Python?
> > > > > 
> > > > > SIP effectively implements its own smartpointer mechanism - so long
> > > > > as the C++ type has a virtual dtor and the C++ object is created
> > > > > from Python.
> > > > 
> > > > Could the SIP documentation be more clear about this? I can only see
> > > > that it says that pointers to C structs will be free'ed, and that C++
> > > > objects are destroyed. I need to make sure that C++ objects are
> > > > deleted, not just destroyed.
> > > 
> > > ...err, what's the difference?
> > 
> > Can you define "destroying" a C++ object as deleting it? When you say
> > destroy, I think of invoking its destructor. A destroyed object isn't
> > necessarily freed, and it's not immediately obvious how to free a
> > heap-allocated object in C++ (depends on how it's allocated).
> 
> Stroustrup talks of objects being "destroyed by delete". As far as SIP is
> concerned they are synonymous.

Sure, delete implies that an object be destroyed, but destruction does
not imply deletion. Look at the first link resulting from googling
"c++ destroy object":
http://publib.boulder.ibm.com/infocenter/lnxpcomp/v8v101/index.jsp?topic=/com.ibm.xlcpp8l.doc/language/ref/cplr380.htm.
 Destroying an object is defined as having its destructor invoked.

You may not agree with me, but your documentation *will* leave things
to be desired, from its audience's perspective, since it doesn't
completely define at least SIP's handling of object ownership.

Arve
_______________________________________________
PyQt mailing list    PyQt@riverbankcomputing.com
http://www.riverbankcomputing.com/mailman/listinfo/pyqt


[prev in list] [next in list] [prev in thread] [next in thread]